Causal Evidence from a Post-Rationing Reform
05 June, 2025
TODO:
It’s a big problem
Limited credible causal evidence
…
A lot of evidence for a detrimental effect of alcohol on health
Difficult to identify causal …
Natural experiments, a fix?
Guidelines:
900 municipalities in 1968, 288 in 1996 (Source: REGINA, SCB)
Mortality data aggregated to 1995 municipal borders
Martality sample restricted to 1968-1996
The population equation of interest is
\[\begin{equation} Y_{mt} = \gamma_m + \lambda_t + \beta_{mt} \text{store}_{mt} + \epsilon_{mt}\, , \label{eq:pop-eq-of-interest} \end{equation}\]
1st store test uses municipalities with no store a the control group
2nd store test uses municipalities with only one store as control group1
1st store test: positive increase1
2nd store test ATT: 33%2
Instead: Callaway and Sant’Anna (2021)
Concern: cross-boundary spillover
Test: define minicipalities with at least one store as treated when the neighboring municipality open a store.
If SUTVA is violated, there should be a negative effect on sales.
There is indeed some violations of the SUTVA assumption (ATT = 5%) for the full sample.
Using the 2nd store test sample, there is no detectable violations agains SUTVA.
Caveats:
If we multiply the coefficient on total alcohol sales from the ‘Saturday’ experiment by the number of days Systembolaget is open (today), we get \(0.035 \times 6 = 0.21\), compared to the \(0.25\) from the 2nd store test (column 5).
….